Please note this is NOT finished, and I do not presently consider it's outputs useful to an end user. It will try to give you feedback, but do not take it's feedback seriously. I told an older version of this thing an amazing STAR story, and due to an error in my prompt, it absolutely shredded my story. If it does this please ignore it, I think your story was probably great. At present I'm building a classifier to allow the S T A R widget to light up based on where you are in your story, expanding Pipecat-AI's support for Azure AI Foundry, and also experimenting with deployment patterns for small models on Azure.
Your data, if I can manage to free up the time, will be used solely by me and my wife (a proper datascientist) to train a text (and potentially acoustic) classfier that makes the widget below light up correctly as you tell your STAR story. That means by using this demo, you agree to share your data with us and it will remain exclusively with us.
My primary goal with this app is to create a voice experience wherein the user interacts with a voice AI, and UI widget is updated throughout the course of the conversation based on the progress of the conversation. So to play around with it, just say the word "situation", "task", "action", or "result", and system should highlight the appropriate phase.